Point Transformer V3: Simpler, Faster, Stronger

By Xiaoyang Wu, Li Jiang, Peng-Shuai Wang, Zhijian Liu, Xihui Liu, Yu Qiao, Wanli Ouyang, Tong He, Hengshuang Zhao. Computer Vision and Pattern Recognition, 2023

958 citations recorded by the source corpus.

This paper is not motivated to seek innovation within the attention mechanism. Instead, it focuses on overcoming the existing trade-offs between accuracy and efficiency within the context of point cloud processing, leveraging the power of scale. Drawing inspiration from recent advances in 3D large-scale representation learning, we recognize that model performance is more influenced by scale than by intricate design. Therefore, we present Point Transformer V3 (PTv3), which prioritizes simplicity and efficiency over the accuracy of certain mechanisms that are minor to the over-all performance after scaling, such as replacing the precise neighbor search by KNN with an efficient serialized neighbor mapping of point clouds organized with specific patterns. This principle enables significant scaling, expanding the receptive field from 16 to 1024 points while remaining efficient (a 3 x increase in processing speed and a 10 x improvement in memory efficiency compared with its pre-decessor, PTv2). PTv3 attains state-of-the-art results on over 20 downstream tasks that span both indoor and out-door scenarios. Further enhanced with multi-dataset joint training, PTv3 pushes these results to a higher level.

Subfields: Computer Vision and Pattern Recognition, Mechanical Engineering

View Paper: https://www.semanticscholar.org/paper/10b238f8b0da6edaa24008448e3f0b7a01d7c268

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

1. Zhikang Zou
2. Xiang Bai
3. Xin Zhou
4. Xiao Tan
5. Xingkui Zhu
6. Xiaoqing Ye
7. Wei Xu
8. Dingkang Liang

Citations in the corpus

• PointMamba: A Simple State Space Model for Point Cloud Analysis (2024, methodology): “We note that some methods [54, 51] utilize space-filling curves to partition the point cloud for capturing spatial contexts, whereas our work mainly focuses on transferring the point clouds to serialization-based sequences and combine with Mamba to implement global modeling.”
• Aligning Cyber Space With Physical World: A Comprehensive Survey on Embodied AI (2024, neutral): “In this section, we will introduce the existing datasets, discusstherelatedmethods,describethemetricsusedtoevaluate modelperformance,andaddresstheremaininglimitationsof thistask.”